Microelectromechanical systems (MEMS) microphones are commonly used in mobile telephones and other consumer electronic devices, embedded systems and other devices. A MEMS microphone typically includes a conductive micromachined diaphragm that vibrates in response to an acoustic signal. The microphone also includes a conductive plate parallel to, and spaced apart from, the diaphragm with air or another dielectric between the conductive plate and the diaphragm. The diaphragm and the conductive plate collectively form a capacitor, and an electrical charge is placed on the capacitor, typically by an associated circuit. The capacitance of the capacitor varies rapidly as the distance between the diaphragm and the plate varies due to the vibration of the diaphragm. Typically, the charge on the capacitor remains essentially constant during these vibrations, so the voltage across the capacitor varies as the capacitance varies.
The varying voltage may be used to drive a circuit, such as an amplifier or an analog-to-digital converter, to which the MEMS microphone is connected. Such a circuit may be implemented as an application-specific integrated circuit (ASIC). A MEMS microphone connected to a circuit signal processing circuit is referred to herein as a “MEMS microphone system” or a “MEMS system.” A MEMS microphone die and its corresponding ASIC are often housed in a common integrated circuit package to keep leads between the microphone and the ASIC as short as possible, such as to avoid parasitic capacitance caused by long leads.
When used in consumer electronics devices and other contexts, MEMS microphone systems may be subjected to widely varying amplitudes of acoustic signals, including background noise. For example, a mobile telephone used outdoors under windy conditions or in a subway station subjects the MEMS microphone to very loud acoustic signals. Even under quite ambient conditions, a user may hold a microphone too close to the user's mouth or speak in too loud a voice for the MEMS microphone system. Under these circumstances, the diaphragm may reach its absolute displacement limit, and the resulting signal may therefore be “clipped,” causing undesirable distortion. Even if the diaphragm does not reach its absolute displacement limit, the ASIC or other processing circuitry may not be able to handle the peaks of the electrical signal from the MEMS microphone, i.e., the processing circuitry may have insufficient “headroom” for the signal from the MEMS microphone, and the signal may be clipped. Clipping can cause a loss of signal contents. For example, if a speech signal is clipped, the output signal waveform becomes flat and no longer varies with the human speech. Thus, during the clipped portion of each cycle, the signal conveys no intelligible content.
On the other hand, if a user speaks too softly, such as in the absence of background noise, the amplitude of the signal produced by the MEMS microphone may be very low, yielding a low signal-to-noise ratio (SNR) in the microphone signal or in a signal produced by some “downstream” circuit. Prior art audio systems sometimes include automatic gain control circuits that vary the gain of amplifiers that process signals produced by microphones. However, such gain adjustments can be abrupt and, thus, audibly displeasing. Furthermore, such gain adjustments do not change the headroom of the amplifiers. Thus, conventional MEMS and other capacitor microphone systems are susceptible to signal distortion and poor signal-to-noise characteristics.